Reamer commented on pull request #3859:
URL: https://github.com/apache/zeppelin/pull/3859#issuecomment-661823077
With the following change, every user should be able to modify Conda
packages. However, I do not recommend this, because it can have side effects if
several users are running
xiejiajun opened a new pull request #3860:
URL: https://github.com/apache/zeppelin/pull/3860
### What is this PR for?
- fix the [CI
failure](https://travis-ci.org/github/apache/zeppelin/builds/709913046) due to
[PR-3852](https://github.com/apache/zeppelin/pull/3852)
### What
I didn't compile it myself, I just use binaries that Jeff created for
preview2. My point is that it worked out of box in preview1, and previous
versions, and should continue be the same, otherwise it's a very breaking
change that requires that people know about that...
On Tue, Jul 21, 2020 at
That's right, In that PR, I exclude hadoop jars from zeppelin distribution,
so that we can support both hadoop2 and hadoop3 (user could set
USE_HADOOP=true in zeppelin-env.sh, so that zeppelin run command `hadoop
classpath` and put all the hadoop jars in classpath of zeppelin.
But for this issue,
Hi Alex,
It seems that Hadoop classes are missing. Do you include Hadoop jars
with "-P include-hadoop"?
I think it's related to
https://github.com/apache/zeppelin/commit/6fa79a9fc743f2b4321ac9e8713b3380bb4d64c9#diff-600376dffeb79835ede4a0b285078036.
Philipp
Am 21.07.20 um 11:28 schrieb
Hi Jeff
I've found another issue in both rc1 & rc2 - if you don't specify the
SPARK_HOME, then the default Spark interpreter doesn't start with following
error if I execute the code for reading from Cassandra:
%spark
import org.apache.spark.sql.cassandra._
val data =
alexott commented on pull request #3859:
URL: https://github.com/apache/zeppelin/pull/3859#issuecomment-661716738
We need to think about target audience in this case. If we're targeting
analysts, they are mostly not the people who are creating Docker images, etc.
Requirement to build
Reamer commented on pull request #3859:
URL: https://github.com/apache/zeppelin/pull/3859#issuecomment-661705487
> Philipp - can we also solve the problem with access rights for Conda
posted by some users on SO: